低排放石油和天然气：实现清洁与可靠的平衡.pdf

1、Transformation Data Management in bp with Unity CatalogDriving Innovation,Governance,and Scalability with DataSrinivas Chandolu&Krishanu RoyJune 20252Our Ecosystem&ValuesHow we got hereChallenges Along the WayHow unity catalog helpsOur Migration JourneyKey TakeawaysAGENDA3informationknowledgeinsight

2、swisdomimpactdataRaw facts and figures without contextExamples:Sensor readings,transaction logs,timestamps,standardsknowledgeOrganized information applied to tasks or decisionsExamples:equipment maintenance schedules,reservoir behavior modelsDeeper understanding derived from patterns or correlations

3、Examples:Identifying declining well performance,predicting equipment failureinsightsinformationProcessed data with contextExamples:Well pressure trends,production rates,daily drilling reportsStrategic judgment based on experience and insightsExamples:field development planningwisdomActionable outcom

4、es influencing business performanceExamples Increased production efficiency,reduced downtime,improved safety complianceimpactData platforms landscape+80B streaming data point+2B streaming tagsData platforms philosophyOur Ecosystem&Values4How we got here2014Formation of on-premise upstream data lake

5、pivotal to sending to external vendors2015-20162017-20182019-20222023 onwardsStreaming setup on TibcoData lake setup in AWSbp decides to move to public cloudMigration of upstream data lake to Azure(Azure Native as compute)Data Mesh/Fabric Vision move to data hubsUpstream data lake morphs into Azure

6、Data LakeDatabricks assessment and full-scale migration to DatabricksCreation of Unified data platform strategyMigration to Unity CatalogFocus of reusability and federation5Complex service dependenciesInconsistent data experienceConfusing choices of competing platforms Data sharing on material copie